ASML gives the world's leading chipmakers the power to mass produce patterns on silicon. Originating from Veldhoven, The Netherlands, ASML is a big employer in the Brainport area. ASML's best known product is the TWINSCAN, a lithography machine used in most of the world's chip production.
Product
Chips produced with TWINSCAN are made in so called wafers. These wafers are printed in layers, with some taking up hours to produce. Mistakes in the production process are costly and unnoticable with the human eye. This is where YieldStar comes in. YieldStar is a machine that is able to measure the wafers on a list of KPI's and determines of the wafer is of sufficient quality. This prevents mistakes in the production process to be noticed too late and allows adjustment early on.
Assignment
Situation
Customers of ASML require new KPI's to be added consistently with advancements in the TWINSCAN machine. The codebase of YieldStar is vast and contains a lot of legacy code. This causes a lot of work when new KPI's have to be added, since a huge amount of small changes have to be made throughout the codebase. With a standard addition of a KPI, it will take one (1) developer an entire sprint (2 weeks) to add a single KPI.
Task
To speed up development time the Reporting Automation Tool (RAT) was envisioned. The RAT generates code within the YieldStar project to assist in the addition of new KPI's and reduce errors during the addition of new KPI's
Action
The RAT was extended by adding support for all major reporting formats used within YieldStar (CSV, ADEL, XML). The YieldStar codebase was also refactored at points where new KPI's had to be added to allow for easier code generation. This also involved the standardisation of most KPI's to allow easier future generation. Lastly the RAT was prepared for and eventually released to the production environment of YieldStar.
Result
Developers at the Reporting team within the YieldStar project now actively use the RAT to speed up the addition of new KPI's to the YieldStar codebase. The use of the RAT has reduced the amount of time it needs to add a new KPI significantly and has also reduced the margin of error when adding new KPI's.
Tools and Methods
Programming Languages
The reporting component of YieldStar is written in C#. This was the obvious choice for the RAT as well, since both projects are heavily intertwined. In addition to the use of C#, WPF was used to deliver a comprehensive interface for developers to use. Furthermore, XML was used to store configuration data for the RAT.
Development Methods
During the project the RAT was embedded into a running reporting scrum team at ASML. Naturally the RAT project adhered to the in-place agile working environment. Thus the RAT project was developed using Agile Scrum. To adhere to the high standards of delivery for ASML products the RAT project was also tested extensivly. This was documented using TPS (Test Performance Specification) and TAR (Test Analysis Report) documents.
